Re: Announcement: Github Actions Self-Hosted Runners Testing

[Elias Segundo Antonio]

Hi Damon

I find the approach quite interesting, and I agree that if the YML workflow
is no longer required when using direct triggers, then Self-Hosted runners
are not necessary for those jobs. Are those workflows planned to be removed
in the future when the triggers are synced?

On Fri, Dec 9, 2022 at 11:03 AM Damon Douglas via dev 
wrote:

> Thank you for this information.  For Playground Backend and Frontend, are
> self-hosted runners no longer needed since we are using Cloud Build?
>
> *For those who would like more explanation, please see below.*
>
> Playground is our https://play.beam.apache.org/ that allows us to execute
> Apache Beam code snippets in a modern browser.  Deployed on Google Cloud,
> we develop automation to build container images and other infrastructure
> tasks of the backend and frontend code as a result of GitHub pull request
> merges.  Because we target Google Cloud, we simplify the automation using a
> Cloud Build trigger [1].  Cloud Build allows this image building and other
> automation tasks as a result of GitHub events such as a pull request or
> merge.  When INFRA-23956 [2] resolves, code in the Beam repository related
> to the Beam playground will send a message to Cloud Build to trigger the
> container image build and other related CI/CD events, hence the name "Cloud
> Build trigger".  Think of Cloud Build as a Google Cloud native GitHub
> action-like service that is appropriate when exclusively targeting Google
> Cloud.
>
> References:
>
> 1. https://cloud.google.com/build/docs/triggers
> 2. https://issues.apache.org/jira/browse/INFRA-23956
>
> On Fri, Dec 9, 2022 at 6:57 AM Elias Segundo Antonio <
> elias.segu...@wizeline.com> wrote:
>
>> Hello Apache Beam Team
>>
>> We’d like to announce that we will start testing GitHub Actions
>> Self-Hosted Runners in the Apache Beam Repository. Infrastructure will be
>> in place today at 9:30 AM UTC-6 to migrate the first set of workflows from
>> GitHub-provided Runners to Self-Hosted (for the time being, only Windows
>> and Linux runners are being replaced).
>>
>> If everything runs smoothly, the following list of workflows is expected
>> to be running in Self-Hosted runners by the end of the day:
>>
>>- Build Playground Backend
>>- Build Playground Frontend
>>- Go Tests
>>- Issue Tagger
>>- Java Tests
>>- Label PRs
>>- Local Env Tests
>>- Report Generator
>>- PR bot PR Updates
>>- PR bot PRs needing attention
>>- PR bot update Reviewers
>>- Self Assign
>>- Stale
>>- Triaged on assign
>>- Typescript tests
>>
>> We will be monitoring the behavior of the runners throughout the day; no
>> downtime for GitHub Actions workflows is expected since the new workflows
>> will be merged incrementally and runners provided by GitHub Actions will
>> continue working as normal. However, please don’t hesitate to report if you
>> encounter any issues related to the previous workflows list while we test
>> the infrastructure.
>>
>> Regards.
>> --
>>
>> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
>> wizeline.com <https://www.wizeline.com/>
>>
>> Site Reliability Engineer I
>>
>> elias.segu...@wizeline.com | +52 712 2722589 <+52%20712%20272%202589> (m)
>>
>> Remote Office, Mexico
>>
>>
>>
>>
>>
>>
>>
>>
>> *This email and its contents (including any attachments) are being sent
>> toyou on the condition of confidentiality and may be protected by
>> legalprivilege. Access to this email by anyone other than the intended
>> recipientis unauthorized. If you are not the intended recipient, please
>> immediatelynotify the sender by replying to this message and delete the
>> materialimmediately from your system. Any further use, dissemination,
>> distributionor reproduction of this email is strictly prohibited. Further,
>> norepresentation is made with respect to any content contained in this
>> email.*
>
>

-- 

Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
wizeline.com <https://www.wizeline.com/>

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Announcement: Github Actions Self-Hosted Runners Testing

[Elias Segundo Antonio]

Hello Apache Beam Team

We’d like to announce that we will start testing GitHub Actions Self-Hosted
Runners in the Apache Beam Repository. Infrastructure will be in place
today at 9:30 AM UTC-6 to migrate the first set of workflows from
GitHub-provided Runners to Self-Hosted (for the time being, only Windows
and Linux runners are being replaced).

If everything runs smoothly, the following list of workflows is expected to
be running in Self-Hosted runners by the end of the day:

   - Build Playground Backend
   - Build Playground Frontend
   - Go Tests
   - Issue Tagger
   - Java Tests
   - Label PRs
   - Local Env Tests
   - Report Generator
   - PR bot PR Updates
   - PR bot PRs needing attention
   - PR bot update Reviewers
   - Self Assign
   - Stale
   - Triaged on assign
   - Typescript tests

We will be monitoring the behavior of the runners throughout the day; no
downtime for GitHub Actions workflows is expected since the new workflows
will be merged incrementally and runners provided by GitHub Actions will
continue working as normal. However, please don’t hesitate to report if you
encounter any issues related to the previous workflows list while we test
the infrastructure.

Regards.
-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Re: Metrics-staging Deletion

[Elias Segundo Antonio]

That's the main suspicion with this cluster, however there is no
Grafana image deployed in metrics-staging, the only one deployment
currently is an old version from gcr.io/apache-beam-testing/metrics,
which is also present in the 'metrics' cluster, because of that we
could only find the DNS and External IP for 'metrics', the service is
not present in 'metrics-staging'.


The second possible scenario is that this cluster is more like a
testing cluster for metrics and not actually a staging version as it
is not actively used.


Probably new versions and changes to the Grafana image are tested
against 'metrics-staging' and all the resources are removed once the
process is finished, this will explain the lack of deployments, in
both cases more information is required to justify the prevalence of
the cluster, a better option would be to recreate the test cluster
each time its used to avoid long idle periods.


Regards.

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Metrics-staging Deletion

[Elias Segundo Antonio]

Hi Apache Beam Team


We have been monitoring the usage metrics of the ‘metrics-staging’ cluster,
the average cpu usage was 7.78% during the last 3 months, a deletion was
requested as it seems that the cluster is not being actively used, also
there are no references to this cluster in the code base.


[image: Pasted Graphic 3.png]



If you have any additional info about this cluster please let us know to
document it and cancel the process, if there is no objection we would
proceed with the deletion.


Thanks in advance.



Regards.

-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Re: Upgrade GKE Clusters Versions

[Elias Segundo Antonio]

Hello Apache Beam Team,


We have finished the upgrade process, both clusters are running now on GKE
1.22 and the deprecated APIs are upgraded, validation tests passed without
further issues.


Also, the InfluxDB Auto Backup in metrics cluster which stopped working on
Jan 27 2022, was fixed after the upgrade performed on March 11 2022.


Regards.

On Wed, Mar 23, 2022 at 1:06 PM Elias Segundo Antonio <
elias.segu...@wizeline.com> wrote:

>
> Hello Apache Beam Team!
>
>
> As part of the work required for BEAM-13058[1] we are going to upgrade the
> Kubernetes APIs version for “io-datastores” & “metrics” clusters,  the
> GKE Version will be updated to 1.22 for both control plane and nodes, so
> you may experiment short downtimes when executing performance tests related
> to that cluster.
>
>
>
> The operations will start today at 1:15 pm CST.
>
>
>
> We will send a second announcement when the process is finished.
>
>
>
> Regards.
>
>
>
> [1] https://issues.apache.org/jira/browse/BEAM-13058
>
>
> --
>
> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
> wizeline.com <https://www.wizeline.com/>
>
> Site Reliability Engineer I
>
> elias.segu...@wizeline.com | +52 712 2722589 (m)
>
> Remote Office, Mexico
>


-- 

Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
wizeline.com <https://www.wizeline.com/>

Site Reliabiity Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Upgrade GKE Clusters Versions

[Elias Segundo Antonio]

Hello Apache Beam Team!


As part of the work required for BEAM-13058[1] we are going to upgrade the
Kubernetes APIs version for “io-datastores” & “metrics” clusters,  the GKE
Version will be updated to 1.22 for both control plane and nodes, so you
may experiment short downtimes when executing performance tests related to
that cluster.



The operations will start today at 1:15 pm CST.



We will send a second announcement when the process is finished.



Regards.



[1] https://issues.apache.org/jira/browse/BEAM-13058


-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

[Question] Upgrade Kubernetes APIs - BEAM-13058

[Elias Segundo Antonio]

Hi! Apache Beam Team,


We are currently working on the task BEAM-13058 [1] and we are upgrading
the Kubernetes API on metrics and io-datastores clusters in order to make
them work in the 1.22 API version.


We already tested the changes in the metrics cluster and we also have
tested the changes for Elasticsearch and Cassandra Kubernetes objects, both
are working smoothly.



We are wondering if you could give us more details regarding how these
ElasticSearch and Cassandra Kubernetes objects [2] are being used and
deployed, so we can test a more precise environment.


[1] https://issues.apache.org/jira/browse/BEAM-13058

[2] https://github.com/apache/beam/tree/master/.test-infra/kubernetes



Thanks in advance.


Regards.

-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Re: Upgrade GKE io-datastores Cluster

[Elias Segundo Antonio]

Hello Apache Beam Team

We have finished the upgrade of the io-datastores cluster, GKE Version is
1.20 now and the default container runtime was changed from Docker to
Containerd. Validation tests passed without further issues.

Regards.

On Mon, Mar 14, 2022 at 12:47 PM Elias Segundo Antonio <
elias.segu...@wizeline.com> wrote:

> Hello Apache Beam Team!
>
>
>
> As part of the work required for BEAM-13798 [1] & BEAM-13059 [2] we are
> going to perform an upgrade plan for the cluster ”io-datastores” , the GKE
> Version and container runtime will be updated for both control plane and
> nodes,  so you may experiment short downtimes when executing performance
> tests related to that cluster.
>
>
>
> The operations will start today at 1:00 pm CST.
>
>
>
> We will send a second announcement when the process is finished.
>
>
>
> Regards.
>
>
>
> [1] https://issues.apache.org/jira/browse/BEAM-13798
> [2] https://issues.apache.org/jira/browse/BEAM-13059
>
> --
>
> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
> wizeline.com <https://www.wizeline.com/>
>
> Site Reliability Engineer I
>
> elias.segu...@wizeline.com | +52 712 2722589 (m)
>
> Remote Office, Mexico
>


-- 

Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
wizeline.com <https://www.wizeline.com/>

Site Reliabiity Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Upgrade GKE io-datastores Cluster

[Elias Segundo Antonio]

Hello Apache Beam Team!



As part of the work required for BEAM-13798 [1] & BEAM-13059 [2] we are
going to perform an upgrade plan for the cluster ”io-datastores” , the GKE
Version and container runtime will be updated for both control plane and
nodes,  so you may experiment short downtimes when executing performance
tests related to that cluster.



The operations will start today at 1:00 pm CST.



We will send a second announcement when the process is finished.



Regards.



[1] https://issues.apache.org/jira/browse/BEAM-13798
[2] https://issues.apache.org/jira/browse/BEAM-13059

-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Re: Upgrade GKE Metrics Cluster

[Elias Segundo Antonio]

Hello Apache Beam Team

We have finished the upgrade of the metrics cluster, GKE Version is 1.20
now and the default container runtime was changed from Docker to
Containerd. The Grafana Dashboard is healthy and the validation tests
passed without further issues. On Monday we will execute a similar plan on
io-datastores and we will let you know the time window for the upgrade in a
new notification.

Regards.

On Fri, Mar 11, 2022 at 9:19 AM Elias Segundo Antonio <
elias.segu...@wizeline.com> wrote:

> Hello Apache Beam Team
>
> Jenkins is back online, so we will continue with the upgrade plan, we will
> let you know when we finish.
>
> Regards.
>
> On Thu, Mar 10, 2022 at 10:54 AM Elias Segundo Antonio <
> elias.segu...@wizeline.com> wrote:
>
>> Hello Apache Beam Team
>>
>> We did the upgrade of the cluster control plane for Metrics
>> from 1.18.20-gke.901 to 1.19.16-gke.3600, there wasn't downtime during this
>> process and we would like to perform some validation tests before moving
>> forward, we noticed that Jenkins CI has availability issues [1] and it is
>> returning a proxy error. We are going to wait until this server is back
>> online to continue with the upgrade plan.
>>
>> Please let us know if you have any questions or comments.
>>
>> [1] https://ci-beam.apache.org/job/beam_PostCommit_BeamMetrics_Publish
>>
>> Regards.
>>
>> On Thu, Mar 10, 2022 at 9:56 AM Elias Segundo Antonio <
>> elias.segu...@wizeline.com> wrote:
>>
>>>
>>> Hello Apache Beam Team!
>>>
>>>
>>> As part of the work required for BEAM-13798 [1] & BEAM-13059 [2] we are
>>> going to perform an upgrade plan for the cluster ”metrics” , the GKE
>>> Version and container runtime will be updated for both control plane and
>>> nodes,  so you may experiment short downtimes (max. 5 min each) when
>>> executing the following tests:
>>>
>>>
>>> *Job Name*
>>>
>>> beam_PreCommit_CommunityMetrics
>>>
>>> beam_PostCommit_BeamMetrics_Publish
>>>
>>> beam_Metrics_Report
>>>
>>> beam_Prober_CommunityMetrics
>>>
>>>
>>> Grafana dashboard may experience the same downtime situation, but
>>> io-datastores will continue working as usual.
>>>
>>>
>>> The operations will start at 10:00 AM CST.
>>>
>>>
>>> We will send a second announcement when the process is finished.
>>>
>>>
>>> Regards.
>>>
>>>
>>> [1] https://issues.apache.org/jira/browse/BEAM-13798
>>>
>>> [2] https://issues.apache.org/jira/browse/BEAM-13059
>>> --
>>>
>>> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
>>> wizeline.com <https://www.wizeline.com/>
>>>
>>> Site Reliability Engineer I
>>>
>>> elias.segu...@wizeline.com | +52 712 2722589 (m)
>>>
>>> Remote Office, Mexico
>>>
>>
>>
>> --
>>
>> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
>> wizeline.com <https://www.wizeline.com/>
>>
>> Site Reliabiity Engineer I
>>
>> elias.segu...@wizeline.com | +52 712 2722589 (m)
>>
>> Remote Office, Mexico
>>
>
>
> --
>
> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
> wizeline.com <https://www.wizeline.com/>
>
> Site Reliabiity Engineer I
>
> elias.segu...@wizeline.com | +52 712 2722589 (m)
>
> Remote Office, Mexico
>


-- 

Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
wizeline.com <https://www.wizeline.com/>

Site Reliabiity Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Re: Upgrade GKE Metrics Cluster

[Elias Segundo Antonio]

Hello Apache Beam Team

Jenkins is back online, so we will continue with the upgrade plan, we will
let you know when we finish.

Regards.

On Thu, Mar 10, 2022 at 10:54 AM Elias Segundo Antonio <
elias.segu...@wizeline.com> wrote:

> Hello Apache Beam Team
>
> We did the upgrade of the cluster control plane for Metrics
> from 1.18.20-gke.901 to 1.19.16-gke.3600, there wasn't downtime during this
> process and we would like to perform some validation tests before moving
> forward, we noticed that Jenkins CI has availability issues [1] and it is
> returning a proxy error. We are going to wait until this server is back
> online to continue with the upgrade plan.
>
> Please let us know if you have any questions or comments.
>
> [1] https://ci-beam.apache.org/job/beam_PostCommit_BeamMetrics_Publish
>
> Regards.
>
> On Thu, Mar 10, 2022 at 9:56 AM Elias Segundo Antonio <
> elias.segu...@wizeline.com> wrote:
>
>>
>> Hello Apache Beam Team!
>>
>>
>> As part of the work required for BEAM-13798 [1] & BEAM-13059 [2] we are
>> going to perform an upgrade plan for the cluster ”metrics” , the GKE
>> Version and container runtime will be updated for both control plane and
>> nodes,  so you may experiment short downtimes (max. 5 min each) when
>> executing the following tests:
>>
>>
>> *Job Name*
>>
>> beam_PreCommit_CommunityMetrics
>>
>> beam_PostCommit_BeamMetrics_Publish
>>
>> beam_Metrics_Report
>>
>> beam_Prober_CommunityMetrics
>>
>>
>> Grafana dashboard may experience the same downtime situation, but
>> io-datastores will continue working as usual.
>>
>>
>> The operations will start at 10:00 AM CST.
>>
>>
>> We will send a second announcement when the process is finished.
>>
>>
>> Regards.
>>
>>
>> [1] https://issues.apache.org/jira/browse/BEAM-13798
>>
>> [2] https://issues.apache.org/jira/browse/BEAM-13059
>> --
>>
>> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
>> wizeline.com <https://www.wizeline.com/>
>>
>> Site Reliability Engineer I
>>
>> elias.segu...@wizeline.com | +52 712 2722589 (m)
>>
>> Remote Office, Mexico
>>
>
>
> --
>
> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
> wizeline.com <https://www.wizeline.com/>
>
> Site Reliabiity Engineer I
>
> elias.segu...@wizeline.com | +52 712 2722589 (m)
>
> Remote Office, Mexico
>


-- 

Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
wizeline.com <https://www.wizeline.com/>

Site Reliabiity Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Re: Upgrade GKE Metrics Cluster

[Elias Segundo Antonio]

Hello Apache Beam Team

We did the upgrade of the cluster control plane for Metrics
from 1.18.20-gke.901 to 1.19.16-gke.3600, there wasn't downtime during this
process and we would like to perform some validation tests before moving
forward, we noticed that Jenkins CI has availability issues [1] and it is
returning a proxy error. We are going to wait until this server is back
online to continue with the upgrade plan.

Please let us know if you have any questions or comments.

[1] https://ci-beam.apache.org/job/beam_PostCommit_BeamMetrics_Publish

Regards.

On Thu, Mar 10, 2022 at 9:56 AM Elias Segundo Antonio <
elias.segu...@wizeline.com> wrote:

>
> Hello Apache Beam Team!
>
>
> As part of the work required for BEAM-13798 [1] & BEAM-13059 [2] we are
> going to perform an upgrade plan for the cluster ”metrics” , the GKE
> Version and container runtime will be updated for both control plane and
> nodes,  so you may experiment short downtimes (max. 5 min each) when
> executing the following tests:
>
>
> *Job Name*
>
> beam_PreCommit_CommunityMetrics
>
> beam_PostCommit_BeamMetrics_Publish
>
> beam_Metrics_Report
>
> beam_Prober_CommunityMetrics
>
>
> Grafana dashboard may experience the same downtime situation, but
> io-datastores will continue working as usual.
>
>
> The operations will start at 10:00 AM CST.
>
>
> We will send a second announcement when the process is finished.
>
>
> Regards.
>
>
> [1] https://issues.apache.org/jira/browse/BEAM-13798
>
> [2] https://issues.apache.org/jira/browse/BEAM-13059
> --
>
> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
> wizeline.com <https://www.wizeline.com/>
>
> Site Reliability Engineer I
>
> elias.segu...@wizeline.com | +52 712 2722589 (m)
>
> Remote Office, Mexico
>


-- 

Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
wizeline.com <https://www.wizeline.com/>

Site Reliabiity Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Upgrade GKE Metrics Cluster

[Elias Segundo Antonio]

Hello Apache Beam Team!


As part of the work required for BEAM-13798 [1] & BEAM-13059 [2] we are
going to perform an upgrade plan for the cluster ”metrics” , the GKE
Version and container runtime will be updated for both control plane and
nodes,  so you may experiment short downtimes (max. 5 min each) when
executing the following tests:


*Job Name*

beam_PreCommit_CommunityMetrics

beam_PostCommit_BeamMetrics_Publish

beam_Metrics_Report

beam_Prober_CommunityMetrics


Grafana dashboard may experience the same downtime situation, but
io-datastores will continue working as usual.


The operations will start at 10:00 AM CST.


We will send a second announcement when the process is finished.


Regards.


[1] https://issues.apache.org/jira/browse/BEAM-13798

[2] https://issues.apache.org/jira/browse/BEAM-13059
-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

GKE Clusters Upgrade Question

[Elias Segundo Antonio]

Hi Apache Beam Team


In order to keep our GKE Clusters up to date, Basic Auth is going to be
disabled for the following clusters:


* io-datastores

* Metrics

* Metrics-staging


As you may know, OAuth is the current preferred method for authentication
for both Google Cloud CLI and GKE. Prior to version 1.19 another
authentication method existed based on username and password (Basic Auth).
The previous listed clusters are running under version 1.18.20 which is
being deprecated soon and disabling the feature is required for the upgrade.


The upgrade process is straightforward, however our main concern is that if
there are any automation processes using Basic auth then those will stop
working after disabling the feature.


We would like to know your thoughts about this change and if you have some
advice or documentation about what are the main functions of the clusters.


Please let us know if you have any comments or questions, thank you.


Regards.

-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Re: Github Actions Windows Self-hosted Runner Question

[Elias Segundo Antonio]

Hello everyone, hope everything is going well.

We did some updates in the VM and we tested all the GitHub Actions
workflows and all of them passed by now, we also found that a load balancer
may not be needed, as Github runners are assigned based on tags and
availability.

We also started testing the Template Image with Instance Groups and some
minor issues appeared , but everything is working now so far, the good news
is that we can register multiple Self-hosted runners on the same VM in
order to optimize resources.

Unless someone disagrees,  the VM approach seems like a good choice for
this task.

Regards.


On Wed, Jan 26, 2022 at 1:38 PM Ahmet Altay  wrote:

> I do not have much experience in this area. And from the silence it is
> probably true for others on the mailing list as well. If I remember
> correctly, our primary goal was to migrate to self hosted runners. I do not
> remember if we agreed to specifically do this on GKE and not on VMs. If
> there was a previous discussion I would appreciate it if someone could
> share a link.
>
> Unless someone disagrees, it seems like you have the most context in this
> area and I would be in favor of going with your suggestion.
>
> Ahmet
>
>
>
> On Fri, Jan 21, 2022 at 2:52 PM Elias Segundo Antonio <
> elias.segu...@wizeline.com> wrote:
>
>> Hi everyone, hope all is well.
>>
>> We are working on BEAM 12812
>> <https://issues.apache.org/jira/browse/BEAM-12812> and we have some
>> questions regarding the Windows Self-hosted runners.
>>
>> The requirement is to migrate the existing runners to GKE and we have
>> been working on a container based approach using the following image
>> mcr.microsoft.com/windows/servercore:ltsc2019, when using Kubernetes
>> some critical features needed for test compatibility are not available on
>> Windows container image, such as Windows Subsystem for Linux or direct
>> installation of VS Build Tools.
>>
>> Due to those container limitations, some tests were rewritten or adapted
>> to run on both Linux and Windows environments, but this is going to be also
>> required when future tests or steps are added.
>>
>> In addition, to build some tests VS Build Tools is required as well as
>> some dependencies that make the image size grow to 20GB easily, which is
>> not an optimal use case for containers.
>>
>> For the above reasons we are wondering if Kubernetes approach is the best
>> one for this case or if we should start thinking about switching to Virtual
>> Machines instead.
>>
>> Today we ran “Python tests” and “Build Python Source distribution and
>> wheels” workflows in a Windows GCE Instance without further changes and we
>> found that almost all the tests are passing and some of them have minor
>> issues compared with K8s ones.
>>
>> In the below table you can find some of the pros and cons of using VMs
>> against K8s.
>>
>> We would like to know your thoughts on this case in order to find the
>> best approach for Windows Self Hosted Runners.
>>
>> Please let us know if you have any comments or questions, thank you.
>> Regards
>> --
>>
>> Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
>> wizeline.com <https://www.wizeline.com/>
>>
>> Site Reliability Engineer I
>>
>> elias.segu...@wizeline.com | +52 712 2722589 <+52%20712%20272%202589> (m)
>>
>> Remote Office, Mexico
>>
>>
>>
>>
>>
>>
>>
>>
>> *This email and its contents (including any attachments) are being sent
>> toyou on the condition of confidentiality and may be protected by
>> legalprivilege. Access to this email by anyone other than the intended
>> recipientis unauthorized. If you are not the intended recipient, please
>> immediatelynotify the sender by replying to this message and delete the
>> materialimmediately from your system. Any further use, dissemination,
>> distributionor reproduction of this email is strictly prohibited. Further,
>> norepresentation is made with respect to any content contained in this
>> email.*
>
>

-- 

Elias Edgardo Segundo Antonio (He/Him) | <https://www.wizeline.com/> |
wizeline.com <https://www.wizeline.com/>

Site Reliabiity Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Github Actions Windows Self-hosted Runner Question

[Elias Segundo Antonio]

Hi everyone, hope all is well.

We are working on BEAM 12812
 and we have some
questions regarding the Windows Self-hosted runners.

The requirement is to migrate the existing runners to GKE and we have been
working on a container based approach using the following image
mcr.microsoft.com/windows/servercore:ltsc2019, when using Kubernetes some
critical features needed for test compatibility are not available on
Windows container image, such as Windows Subsystem for Linux or direct
installation of VS Build Tools.

Due to those container limitations, some tests were rewritten or adapted to
run on both Linux and Windows environments, but this is going to be also
required when future tests or steps are added.

In addition, to build some tests VS Build Tools is required as well as some
dependencies that make the image size grow to 20GB easily, which is not an
optimal use case for containers.

For the above reasons we are wondering if Kubernetes approach is the best
one for this case or if we should start thinking about switching to Virtual
Machines instead.

Today we ran “Python tests” and “Build Python Source distribution and
wheels” workflows in a Windows GCE Instance without further changes and we
found that almost all the tests are passing and some of them have minor
issues compared with K8s ones.

In the below table you can find some of the pros and cons of using VMs
against K8s.

We would like to know your thoughts on this case in order to find the best
approach for Windows Self Hosted Runners.

Please let us know if you have any comments or questions, thank you.
Regards
-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliability Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*

Contributor permission for Beam Jira Tickets

[Elias Segundo Antonio]

Hello everyone,

Hope you are doing well.

I'm Elias Segundo, and I'm currently working at Wizeline. I would like to
be added as a contributor in the Beam Jira issue tracker to assign myself
to a couple of Beam tasks.

My JiraID is: EliasSegundo

Thank you in advance!

Regards,

-- 

Elias Edgardo Segundo Antonio (He/Him) |  |
wizeline.com 

Site Reliabiity Engineer I

elias.segu...@wizeline.com | +52 712 2722589 (m)

Remote Office, Mexico

-- 
*This email and its contents (including any attachments) are being sent to
you on the condition of confidentiality and may be protected by legal
privilege. Access to this email by anyone other than the intended recipient
is unauthorized. If you are not the intended recipient, please immediately
notify the sender by replying to this message and delete the material
immediately from your system. Any further use, dissemination, distribution
or reproduction of this email is strictly prohibited. Further, no
representation is made with respect to any content contained in this email.*